Noise-induced hearing loss (NIHL) is among the most common occupational afflictions, especially to construction workers. Hearing loss usually progresses unnoticed until it begins to interfere with communication, decreasing in quality of life and posing a serious safety hazard. Precise exposure-response relationships for noise-induced hearing loss, especially for highly variable noise exposures, as found in construction, are lacking. In recent years, the potential for distortion product otoacoustic emissions, (DPOAEs) as a screening tool for early hearing damage, and possibly as a marker of susceptibility for hearing loss has been recognized. However, no prospective studies of DPOAEs in relation to well-characterized noise exposure and standard audiometry have been conducted. We propose to monitor noise exposure in a cohort of newly-hired construction apprentices and controls and characterize the effects of this exposure on hearing acuity (via standard audiometry) and DPOAEs over a four year period. Four hundred construction apprentices and 100 medical students will be recruited into the study. Each subject will be given an audiometric exam and DPOAE measurements every six months for four years. Baseline and follow-up questionnaires will be used to characterize other risk factors for hearing loss, non- occupational exposure to noise, characteristics of work and use of hearing protective devices (HPDs). Noise exposure will be monitored twice yearly on each subject using data-logging noise dosimeters in conjunction with time/activity cards. The dosimeters will collect noise levels using 3 and 5 dB exchange rates as recommended by NIOSH and OSHA, respectively, as well as peak exposure information. An activity-exposure matrix will be developed from these data to provide estimates of average, peak and variability of exposure during work and home activities. These data will be used to estimate individual exposures (using several metrics) over time. Analyses will be conducted to evaluate the relationships between noise exposure (using both average levels and variable exposure metrics) and both audiometric changes and DPOAEs while controlling for covariates. The relationship between audiometric changes and DPOAEs will also be assessed, while controlling for noise exposure metrics and covariates. The study will exploit the repeated measures design on a naive cohort, and individual measures of exposure to avoid and control problems of bias and exposure measurement error.